1. Field of the Invention
This invention relates to a spectrum analyzer and, more particularly, to an improvement of a spectrum analyzer for analyzing an RF frequency or optical signal.
2. Description of the Related Art
As is well known, a spectrum analyzer measures the characteristics of a high-frequency signal within a relatively wide range; for example, within several tens of kHz to several hundred GHz, and displays the spectrum intensity at each frequency on a CRT display section, as a function of frequency. As is shown in FIG. 10, such a spectrum analyzer includes, for example, an RF section (high-frequency circuit) 1, an IF section (intermediate-frequency circuit) 2, a detecting section 3, an A/D converting section 4, a digital memory 5, a CRT display section 6, a sweep signal generating section 7, a data processing and control section 8, and the like.
With the above circuit arrangement, a high-frequency input signal to-be-measured a, input from an input terminal 9, is input to the RF section 1 consisting of a mixer and a local oscillator. In the RF section 1, the input signal a is mixed with a local signal whose frequency is changed in accordance with the signal level of a sweep signal b output from the sweep signal generating section 7. Thus, the signal a is frequency-converted into an intermediate-frequency signal c which is output from the RF section 1 and input to the IF section 2 which incorporates a bandpass filter (BPF). Only a frequency component which coincides with the pass frequency of the bandpass filter passes through the IF section 2, is input to the detecting section 3, and output as a DC detection signal d corresponding in strength to the magnitude of the input frequency component. The detection signal d is then converted into digital data by the A/D converting section 4, in response to the period of a sampling signal e output from the data processing and control section 8 consisting of, for example, a microprocessor. Thereafter, the converted data is stored in the digital memory 5 consisting of, for example, a RAM.
In response to a read signal f output at a predetermined period from the data procesing and control section 8, digital data of the detection signal d for each sampling period, stored in the digital memory 5, is read out in a predetermined order, and supplied to the display section 6. The display section 6 includes an image memory for one frame which can be displayed on a display screen at one time. After sequentially input data values are stored in the image memory, image data read out from the image memory for one frame is displayed on the display screen. Thus, spectrum distribution data is displayed on the display screen of the display section 6, as shown in FIG. 11.
Each spectrum shown in FIG. 11 has an ideal linear shape. However, each spectrum having a width corresponding to a bandwidth of the bandpass filter in the IF section 2 is actually displayed.
The above-mentioned data processing and control section 8 controls the sweep interval and sweep speed of the sweep signal generating section 7, outputs the sampling signal e to the A/D converting section 4, and outputs the read signal f to the digital memory 5, so that spectrum distribution data as shown in FIG. 11 is displayed on the display section 6. In addition, the data processing and control section 8 executes a calculation required for display, on the display section 6, of the spectrum value corresponding to the desired frequency value.
The spectrum analyzer having the above arrangement shown in FIG. 10 has, however, the following problems. In order to examine the characteristics of signal a, accurate distribution data relating to the spectrum values at every frequency contained in the signal a must be obtained, and the total value of the spectrum values, i.e., the power value of the input signal to be measured a must be simultaneously obtained, in many cases.
When the input signal a contains spectrum components distributed over the bandwidth of the bandpass filter in the IF section 2, its power value is measured, as follows, using the conventional spectrum analyzer as shown in FIG. 10. Data values obtained for every period of the sampling signal e in the A/D converting section 4 during a period from the start to the finish of one sweep signal b can be integrated by, for example, the data processing and control section 8. Accurate integration of the power value of the input signal a is based on the assumption that the spectrum values within the entire frequency range of signal a are accurately measured. However, it is generally difficult to predict in advance the distribution of the spectrum components contained in the input signal to be measured. In addition, as described above, when a deviation of the spectrum to be measured, corresponding to the bandwidth of the bandpass filter in the IF section 2, is corrected and the integration is performed, complicated correcting calculation is required. Therefore, it is practically impossible to accurately measure the power value by the above integration method.
In such a spectrum analyzer, only signal components within the specific frequency sweep range selected and set by the data processing and control section 8 are displayed on the display section 6. Actually, high-level spectrum values (frequency components) are often present in a frequency region outside the frequency sweep range. Thus, an operator is often unaware of the high-level frequency components. In addition, saturation in the RF section 1 is neglected, and many measurement errors may be caused. Furthermore, the RF section 1 may be damaged due to excessive input. In particular, when a signal such as a noise signal and an RF pulse signal wherein spectra are distributed in a wide range is measured, the power value and pulse peak power are further increased as compared with each spectrum value. Therefore, the above-mentioned problems tend to occur.
Even if all the spectra are present in the pass band of the bandpass filter in the IF section 2, as in the case of continuous-wave (CW) signal and the power value can be measured in principle, a considerable error is usually generated because the signal to be measured passes through the complicated signal circuits 7 in the RF and IF sections 1 and 2. Therefore, in an attempt to accurately measure the power value of the continuous-wave signal, a signal having well-known accurate power value is input in advance as a reference signal, and an indication value of the spectrum analyzer is calibrated for the frequency to be used. However, the calibration process is quite time-consuming and the efficiency of measurement operation is considerably degraded. In practice, it is almost impossible always to provide a signal source having a well-known accurate power value at a measurement frequency, and the power value must be measured with an insufficient measurement precision.